"=u

PROOFOF THE STRUCTURE OF RETRONECINE

Feb., 1944

[CONTRIBUTION FROM THE NOYES CHEMICAL

LABORATORY, UNIVERSITY OF

257

ILLINOIS]

Structure of Monocrotaline. XI. Proof of the Structure of Retronecine' BY ROGERADAMSAND N. J. LEONARD Retronecine, the base obtained dong with carbon dioxide and monocrotic acid upon alkaline hydrolysis of monocrotaline, has been postulated2** as having structure 1. The presence of the

I

-CH,OH group in the 1-position of the pyrrolizidine nucleus* and the location of the double bond in the l , % p ~ s i t i ~have n ~ been proved conclusively. The location of the secondary hydroxyl group in retronecine (I) or in retronecanol (II)* has been in doubt. The experimental evidence favors either the 6- or the 7-position. The degradation of retronecanol by bromocyanogen indicates that the 7-position is the more likely.e Corroborative evidence for the 7-position is obtained by construction of models representing anhydroplatynecine (111). When the secondary

\"

LO I

I% + \N'co I

Brz

fPClS

co I

r-O--CHz H O { M

W

IX

X

XI

I11

I1

gave 8-methyl-6-m-nitrobenzoylaminovaleric acid (VIII), which would be formed by cleavage of either ring carbon-nitrogen bond. Bromination of this acid (VIII) in the presence of red phosphorus resulted in the formation of two products IX and X. Compound X was identified as N-mnitrobenzoyl-y-methyl- &p-dibromo-a-piperidone by bromination of N-m-nitrobenzoyl-y-methyl-apiperidone (XI) .8 The structure of compound IX was accepted on the basis of the results of the researches of Fischer and Zemplh,' who prepared IV -CHs the analogous compound through a similar series of reactions from pyridine. retronecanol with aluminum t-butoxide. It has &-Bromo-8-methyl-6-m- nitrobenzoylaminovathus been proved that retronecine and its derivatives have the secondary hydroxyl group in the 7- leric acid (IX) was not obtained in a pure state but the crude product was converted to ethyl position. 4-Methylpyridine (V) was reduced catalytically 3-methylpyrrolidine-2-carboxylate(XIV) through to 4-methylpiperidine (VI), which was converted the intermediates XI1 and XIII. The formation into 1-m-nitrobenzoyl-4-methyl-piperidine (VII) of 1-m-nitrobenzoyl-3 -methylpyrrolidine-2-carby a Schotten-Baumann reaction with m-nitro- boxylic acid (XII) resulted from the reaction of benzoyl chloride. Oxidation of compound VI1 alkali upon IX, and 3-methylpyr~olidine-2-carwith potassium permanganate according to the boxylic acid hydrochloride (XIII) was obtained method of &hotten616 and Fischer and Zemp16n7 by the hydrolysis of XI1 with hydrochloric acid. Owing to the difEculty in purifying either XI1 or (I) For previous paper see Adams and Wilkinaon, Tars JOUXNAL. XIII, compound XI11 was esterified directly to 66, 2203 (1943). XIV, which was obtained readily in pure form by (2) Adams, Carmack and Mahan, ibid., 64,2693 (1942). distillation. The physical properties and analyses ( 3 ) Adams and Hamlin, ibid,, 64, 2697 (1942). (4) Adams and Mahan, ibid., W, 2009 (1943). of XIV and its picrate indicated conclusively that (a) Schotten, Ecr., lT, 2644 (1884). the compound was ethyl 3-methylpyrrolidine-2(6) Schotten, ibid., ll, 2385 (1888).

hydroxyl group is in the 7-position, the structural model of anhydroplatynecine involves somewhat less strain. An optically active form of 1-methyl-7-ketopyrrdizidine has now been synthesized by an unequivocal method and found to be identical with the retronecanone (IV)* obtained by oxidation of

"=u

(7') F k h a and Zemplin, ibid.. U , 2069 (1909).

(8) Schniepp and Marvel, T ~ lJomwlt, l 61, 2822 (1939).

Vol. 66

ROGERADAMSAND N. J. LEONARD

258

carboxylate. Proline has been made previously This proved to be identical in melting point, by a similar series of reactions.’ Ethyl 3-methyl- crystalline form as observed under the micropyrrolidine-2-carboxylateunderwent practically scope, and specific rotation with retronecanoae quantitative addition to ethyl acrylate to give oxime, and the melting points of mixtures of the two compounds showed no depression. The picethyl ~-N-(3-methyl-2-carbethoxypyrrolidyl)-propionate (XV), which was cyclized with powdered rates of the aminoketoximes obtained from potassium. The intermediate keto-ester was not natural and synthetic sources were also found to be identical and showed no depression of meltingCHs , -CH3 decomposition point when mixed. Likewise, the IX *I (A-COO. --* ‘ 2 - C O O H --+ picrolonates of the oximes were identical. Since the keto group in retronecanone occupies the same I H.HC1 CO psition as the secondary hydroxyl group in its I XI11 parent, retronecanol, this hydroxyl group has been located definitely at the 7-position, and the precise structure of retronecanol may now be represented by XVII. Retronecine, therefore, has V\NOL structure XVIII; platyne~ine,~ structure XIX; XI1 and des. > ptronecine,’O structure XX. Owing Ho

I

II

CHiOH

W! XVIII

XIS‘

w XI#

isolated but was hydrolyzed and decarboxylated to 1-methyl-7-ketopyrrolizidine(XVI). There may be formed in the ring closure two keto-esters depending upon which ethoxy group is eliminated as ethanol. Both of these, however, on hydrolysis and decarboxylation would yield the same ketone. The picrate, methiodide and oxime of compound XVI were prepared. Although theoretically two racemic modificationsmight be present in the liquid XVI as synthesized only single derivatives were isolated in each case. An attempt was made to isolate an optically active l-methyl-7ketopyrrolizidine from the possible mixture of four forms by the use of Z-m&ydrazide. A pure I-menthydrazone was obtained but it proved to have a different melting point and specific rotation from the 1-menthydrazone of retronecanone. Moreover, hydrolysis of the 1-methyl-7-ketopyrrolizidine 1-menthydrazone and conversion of the ketone to the oxime gave an optically inactive product. In order to decrease the stereochemical possibilities in the final product as a result of working with racemic modifications, a similar series of reactions was carried out using l-j3-methyl-d-mnitrobenzoylaminovaleric acid. This optically active acid was obtained by resolution through the quinidine salt of the racemic acid (VIII). There was thus obtained optically active l-methyl-7ketopyrrolizidine (XVI) which could not consist of a mixture of more than two forms. Like the product from the racemic acid this was an oil. When treated with hydroxylamine, however, a single, pure, optically active oxime was isolated.

I

w xx

to the fact that heliotridine and oxyheliotridane” are merely stereoisomers, respectively, of retronecine and retronecanol, it is also established that these molecules have the secondary hydroxyl group in the 7-position. Monocrotaline, the acetyl and benzoyl esters of retronecine, and retronecine itself have been reduced to r e t r o n e c a n ~ l . ~These ~ ~ ~ reductions ~~~~ have been carried out in solutions of widely different pH over palladium or platinum oxide catalysts, and retronecanol is the only l-methyl-7hydroxypyrrolizidine isomer which has been obtained, always in essentially quantitative yield. Likewise, the catalytic reduction of desoxyretronecine in hydrochloric acid resulted in almost quantitative production of retronecanol, lo and isoheliotridene was reduced in ethanol over platinum oxide to heli~tridane,~ which is related to retronecanol through chlor~retronecane,~~ Reduction of all of these A1I2-unsaturated pyrrolizidines makes carbon atom-1 asymmetrical. There is the possibility of reduction proceeding t o give a mixture of isomers in which the C1-CH3 would be both cis and trans with respect to the Cs-H; the other possibility would be asymmetric reduction proceeding to give either the cis- or the truns-C1-CHa configuration. Asymmetric reduction is indicated because of the fact that synthetic (and, therefore, naturally-obtained) retronecanone possesses the CI-CHI fixed in one stereo(9) Orekhov and Tiedebel, Ber., 68, 650 (1935). (10) Adams and Rogers, TEIS JOURNAL, 68,537 (1941). (11) Menshikov, Bcr., 68, 1051 (1935). (12) Barger, Seshadri, Watt and Yabuta, J . Ckrm. Soc., 11 (1935). (13) Adams and Rogers, Tms JOURNAL, 6i, 2815 (1939). (14) Adams and Rogers. ibid., U, 228 (1941).

Feb., 1944

PROOF OF THE STRUCTURR OF RETRONECINE

259

chemical configuration. This may be deduced crystalline product which separated was collected on a filter; extraction of the acidic filtrate with ether furnished from the fact that the starting material for its additional product. The total combined yield was 340 g. preparation was I-B-methyl-6-m-nitrobenzoyl-(57%) which was purified by recrystallization from dilute aminovaleric acid, in which the Ci-CH, had a fixed acetic acid or aqueous ethanol; colorless elongated prisms, m. D. 103-105'. configuration. Anal. Calcd. for CItHlt.NnO5: C, 55.71; H, 5.75; N, Reduction of retronecanone could yield a mixture of a maximum of four stereoisomeric 1- 9.99. Found: C, 55.78; H, 5.93; N, 9.84. Resolution of dl-,9-Methyl-S-m-nitrobenzoylaminovamethyl - 7 - hydroxypyrrolizidines if C-8 were leric Acid.-A mixture of 100 g. of the dl-acid and 116 g. of racemized in the aluminum t-butoxide oxidation anhydrous quinidine was dissolved in hot absolute ethanol, of retronecanol to retronecanone. A maximum of boiled with Darco and Norite and filtered. The ethanolic two isomers, of which one 'would be retronecanol, filtrate, which amounted t o 450 cc., was diluted with 1300 cc. of anhydrous ether and allowed to stand at room temcould result on reduction of retronecanone if the perature for several days. Rosets of colorless elongated latter possessed only one stereochemical con- prisms separated and these were isolated by filtration, figuration about (2-8. Neither retronecanol nor its washed with anhydrous ether and dried (Fraction 1) ; m . derivatives could be obtained by reduction of p., 125.0-126.5' (cor.); yield 77 g. (36% of total). Rotation. 0.2286 g. made up to 5 cc. with absolute retronecanone over platinum oxide in ethanol or a t 28' gave LYD f5.10"; I , 1; [ a I z 8 D +111.5' in ethanolic hydrochloric acid solution. The re- ethanol (t0.5'). Further recrystallization from absolute ethanolaction products from reductions carried out both anhydrous ether altered neither the melting point nor the in neutral and acid solution were mixtures. In rotat ion. After removal of 77 g. of the most iiisoluble quinidine each of the two experiments it was possible to iso(Fraction I), the filtrate was diluted further with late a pure compound as a picrate, but these were salt anhydrous ether and allowed to stand in the cold. A not identical with each other or with retronecanol further 48 g. (22%) of yellow, semicrystalline material picrate. Upon acetylation of each reduction separated, leaving approximately 42% of the total salt in product mixture and then treatment with methyl solution (Fraction 2). The more soluble salt could not be crystalline. iodide, the same pure acetyl methiodide was iso- obtained I-8-Methylb-m-nitrobenzoylaminovaleric Acid.-Fraclated in small yield. This methiodide was not tion 1 (75 9.) was shaken thoroughly with 500 cc. of 2 N identical but stereoisomeric with acetylretrone- sulfuric acid and 200 cc. of ether, then separated. The canol methiodide. As a consequence of these re- mixture of dilute acid and undissolved quinidine salt was with 200-cc. portions of ether until no more solid sults no attempt was made to synthesize natural shaken material remained. The ethereal extracts were combined. retronecanol from synthetic optically active 1- dried over anhydrous magnesium sulfate, and evaporated methyl-7-ketopyrrolizidine. to give 30 g. of crystalline l-acid. This was purified by recrystallization from aqueous ethanol : colorless rods, Experimental m. p. 113-114'. 4-Methylpiperidine (VI) .-Hydrogenation of 482 g. (5.18 moles) of redistilled 4-methylpyridine (V) (b. p. 142-144") was concluded after twenty-four hours a t 210" and 150-300 atm. with 50 g. of Raney nickel catalyst. The product was fractionally distilled and the fraction boiling at 12G129" was collected16; n 2 0 1.4487; ~ yield 283 g. (55%, or 90% on the basis of unrecovered 4-methylpyridine). l-m-Nitrobenzoyl-4-methylpiperidine(VII).-To a wellstirred mixture of 265 g. (2.67 moles) of 4-methylpiperidine and 120 g. (3.00 moles) of sodium hydroxide in 630 cc. of water was added slowly, during the course of two hours, 500 g. (2.69 moles) of molten m-nitrobenzoyl chloride.6 The temperature was maintained a t 35-40'. When addition was complete the mixture was stirred for an additional hour and then cooled in ice. The yellow crystals were filtered, washed with water and dried; yield of crude product, 582 g. (88%). After recrystallization from aqueous ethanol, 550 g. of large colorless prisms was obtained, m. p. 72-73". The compound may be recrystallized also from benzene-ligroin (b. p. 90-1l0"). Anal. Calcd. for ClsHlsNzOa: C, 62.89; H, 6.49. Found: C, 63.00; H , 6.50. dl-,9-Methylb-m-nitrobenzoylaminovalericAcid (VIII), -A total of 520 g. (2.1 moles) of l-m-nitrobenzoyl-4methylpiperidine was oxidized in thirteen batches of 40 g. each. The amide (40 9.) was dissolved in 2.5 liters of boiling water and a solution of 60 g. of potassium permanganate in 500 cc. of hot water was added during one hour, while stirring vigorously and refluxing. The reaction mixture was filtered hot and the alkaline filtrate was cooled and extracted with ether. The combined ether extracts of thirteen runs furnished 88 g. of crude starting material which again was subjected t o permanganate oxidation. The aqueous solution (above) was acidified with dilute sulfuric acid and cooled in ice. The colorless (16) Wenburg. Aam.. t47, 1 (1888)

Anal. Calcd. for ClaHl~NzOs: C, 55.71; H, 5.75; N o 9.99. Found: C, 55.96; H, 6.02; N,9.92. Rotation. 0.2040 g. made up to 5 cc. with absolute ethanol a t 30' gave a~ -0.205'; I , 1; [ C Y ] " ~ D-5.0' (*0.2'). d-,9-Methyl-S-m-nitrobenzoylaminovaleric Acid.-The ethereal solution containing Fraction 2 was shaken with 2 N sulfuric acid, and 34 g. of the d-acid was obtained by the same procedure as that outlined above. Purificatioii was accomplished by recrystallization from chloroformligroin (b. p. 90-110'): colorless elongated prisms, m. p., 113-114'. Anal. Calcd. for CIYH16N20h: C, 55.71; H, 5.75; N. 9.99. Found: C, 55.94; H, 5.84; N, 10.09. Rotalion. 0.2164 g. made up "to 5 cc. with absolutz ethanol a t 25' gave a D $0.23 ; E, 1 ; [ c Y ] ~ ~ D+5.3 (h0.2'). Bromination of dZ-8-Methyl-S-m-nitrobenzoyl~inovaleric Acid.-An intimate mixture of 50 g. (0.18 mole) of dl-8-methyl-8-m-nitrobenzoylaminovaleric acid and 8 g . of red phosphorus was protected from moisture, cooled in ice. and stirred manually during the thirty-minute addition of 51.5 cc. (ca. 0.94 mole) of dry bromine. When addition was complete the mixture was stirred at room temperature for thirty minutes, warmed t o 90" during ten minutes, and kept at 90' for a further ten minutes. The reaction mixture was cooled in ice and ice-water was added gradually. Excess bromine was reduced by passing a slow stream of sulfur dioxide into the cold solution. After addition of excess sodium carbonate solution the mixture was stirred vigorously and filtered. The sodium carbonate-insoluble material weighed 15 g. It was purified by recrystallization from ethanol, from which it separated in leaf-like aggregates of colorless elongated prisms, m. p. 152-153' (cor.). It proved t o be N-m-nitrobenzoyl-y-methyl-B,B-dihromoe-piperidone (X).

ROGERADAMSAND N. J. LEONARD

260

Vol. 66

rolidine-2-carbo lic acid in 1 1. of 3 N hydrochloric acid was refluxed f o r x u hours. The solution was cooled, and the m-nitrobenzoic acid was removed by filtration. The filtrate was concentrated t o small volume in mcuo, cooled and filtered free of additional precipitated m-nitrobenzoic acid. This filtrate was extracted with three portions of ether and evaporated t o dryness in vacuo t o yield 9 g. of crude 3-methylpyrrolidine-2-carboxylic acid hydrochloride (XIIT) as a viscous oil. This amino acid hydrochloride was esterified by a method analogous t o that used for preparing the ethyl ester of l-proline.l@ The crude material N-m-Nitrobenzoylr-methyl-o-piperidone(a) .-A so- wa5 suspended in 150 cc. of freshly distilled absolute lution of 2 g. of dl-/3-methylb-m-nitrobenzoylaminovaleric ethanol. The suspension was saturated with dry hydrogen acid (VIII)o in 25 cc. of acetic anhydride was refluxed for chloride which brought about complete solution. The two and one-half hours in all-glass apparatus. The acetic solution was refluxed gently for ten minutes, allowed to anhydride was removed in vacuo and the residue was re- stand for several hours protected from moisture, and then crystallized from aqyeous ethanol: colorless elongated evaporated in vacuo. The residue was redissolved in 100 prisms, m. p. 102-103 . The melting point of a mixture of cc. of absolute ethanol and again saturated with dry this product and the starting material (m. p. 103-104") hydrogen chloride, with natural r d u x h g . After standmg overnight the solution was evaporated t o dryness in vacuo. was depressed t o 85-95'. Anal. Calcd. for C U H I ~ N ~ OC,~ 59.53; : H, 5.38; N, The brown viscous residue was dissolved in 25 cc. of warm absolute eth mol and 175 cc. of anhydrous ether was added. 10.68. Found: C, 59.71; H , 5.44; N, 10.82. The.solutio, was cooled in a n ice-salt-bath and dry amN-m-Nitrobenzoylq-methyl-@, 8-dibromo-a-piperidone was p: s e d through, slowly at first, then in a rapid (X).-To 0.25 g. of N-m-nitrobenzoyl-ymethyl-a-piperi- monla stream until the solution was saturated. It was allowed done in 5 cc. of dry chloroform were added three drops of to stand overnight and, after filtration and removal of the phosphorus trichloride and 0.6 g. of bromine.8 The ether, the residue was distilled under diminished pressure; solution was illuminated by a bright electric light for b. p. 90-91.5" (17.5mm.), 98-97' (21 mm.); ~ B 1.4470; D forty-eight hours, then the reaction mixture was treated d B c 0.993. The yield was 4 g., or an over-all yield of 14% with water and sufficient sodium bisulfite to decompose of the theoretical amount, based upon 50 g. (0.18mole) of the excess bromine. The chloroform layer was separated dE-6-methyl-S-m-nitrobenzoylaminovaleric acid (VIII). and evaporated t o dryness, leaving 0.10 g. of product A n d . Calcd. for C~H,SNO*: C, 61.13; H,9.61; N, which was recrystallized from ethanol: colorless elongated prisms, m. p., 152-153' (cor.). When this product was 8.91;MRD,42.36. Found: C, 61.18; H,9.00;N,8.92; mixed with the alkali-insoluble material obtained iii MRD,42.31. dl-8-methyl-6-m-nitrobenzoylaminovaleric Picrate.-Prepared in and recrystallized from dry benbrominating Lene, the picrate formed long yellow prisms, m. p. 112.5acid, there was no depression of melting point. Bromination of Z-@-Methyl~-m-nitrobenzoylamiaova- 114O. leric Acid-The bromination was carried out in a manner Anal. Calcd. for C I ~ H I ~ N ~C, O ~43.53; : H, 4.69. identical with that used for the dl-acid. The quantities Found: C, 43.67; H , 4.81. employed were 25 g. (0.09 mole) of l-acid, 3.5 g. of red Optically Active Ethyl 3-Metbylpyrrolidine-2-carboxyphosphorus, and 20 cc. of bromine (ca. 0.36mole), and two late.-The same procedure was followed as for the optically products were obtained. The alkali-insoluble bromination inactive compound. The hydrolysis was carried out by product was recrystallized from ethanol: colorless prisms, refluxing a solution of 14 g. of crude optically active l-min. p., 167-168' (cor.). This was I-N-m-nitrobenzoyl-y- iiitrobenzoyl-3-methylpyrrolidine-2-carboxylic acid in 400 methyl-@,@-dibromo-a-piperidone. cc. of 3 N hydrochloric acid for five hours. The crude Anal. Calcd. for ClsHl~BraNaO~: C, 37.17; H, 2.88; optically active 3-methylpyrrolidine-2-csrboxylic acid X, 6.67. Found: (2.37.33;H.3.05; N, 6.54. hydrochloride weighed 4.7 g. and was esterified with Rotation. 0.1211 g. made up t o 5 cc. with pyridine a t absolute ethanol saturated with dry hydrogen chloride as described above. The ethyl ester boiled at 97-98', (23 30" gave a~ -0.49"; I, 1; [ a ] " -20.2" ~ (*0.2'). The alkali-soluble product of bromination weighed 19 g. mrn.); +D 1.4458; yield, 1.2 g. but was not obtained crystalline. It was crude optically Anal. Calcd. for QHlsNOn: N, 8.91. Found: N, a-bromo-/3-methyl-6-m-nitrobenzoylaminovaleric8.76. active acid and was used directly in the next step. Rotation. 0.0686 g. made up to 5 cc. with absolute l - m - N i t r o b e n z o y l - 3 - m e t h y l ~ o l i ~ e - Z ~ ~ xAcid y l i c ethanol at 28" showed no measurable rotation. (=I).-A solution of 42 g. of crude a-bromo-&methylE t h y l b N - ( 3-Methyl-Z-carbethospyrrolidyl -propio6-rn-nitrobenzoylaminovalericacid (IX) in 350 cc. of 1 K nab (XV).ll-A solution of 3.8 g. (24 mdlimoles) o ethyl 3sodium hydroxide was kept a t 37' for forty-eight hours. methylpyrrolidine-2-carboxylateand 8 cc. (ca. 70 milliThe red solution was nearly neutralized with 1 N sulfuric moles) of ethyl acrylate containing a trace of hydroquinone acid and concentrated in vacuo t o 100 cc. The solution was heated under reflux for twenty-four hours on the steamwas cooled, acidified strongly with sulfuric acid and exprotected from moisture. The clear yellow solution tracted with five 100-cc. portions of ether. The combined bath, was fractionally distilled in vacuo. The boiling point of cther extracts were washed with a small quantity of water, colorless product was 163.5-165.5' (18 mm.), 126dried over anhydrous magnesium sulfate, and the ether the 127' (4mm.); n% 1.4523; dm4 1.032; yield, 6.0 g. (97%). was removed. The residue amounted t o 29 g. of thick Anal. Calcd. for ClsH~*NO4: C, 60.69; H, 8.98; N, yellow oil which could not be induced to crystallize. 5.44; MRD, 07.46. Found: C,60.73; H, 8.96; N, 5.36; The crude l-m-nitrobenzoyl-3-methylpyrrolidine-2-carMRD, 67.27. boxylic acid was esterified directly. Optically Active l-~~itrobenzoyl-3-methplpyrrdidlne- Picrate.-Repared in anhydrous ether and recrystallized 2-carborgiic Acid.--This compound was prepered in the from dry benzene, the picrate formed glistening yellow same manner as the above acid, starting with a solution of platelets, m. p., 9 8 - 9 9 O . 19 g. of crude optically active a-bromo-8-methyl-8-mA d . Calcd. for C1gHtJ\T4011: C, 46.92; H, 5.39. nitrobenzoylaminovaleric acid in 160 cc. of 1 N sodium Found: C,47.07; H , 5.50. hydroxide. The product was obtained as 14 g. of yellow (16) Rapfhamma and Matthes, Z.ghysiol. Chcm., %St, 43 (1983). which was used directly in the next step in the synthe-

Anal. Calcd. for Cl,Hl~Br~Nn04:C, 37.17; H, 2.88; N, 6.67. Found: C,37.36; H , 3.08;N,6.62. After removal of the alkali-insoluble material the filtrate was cooled, acidified with 1 N sulfuric acid and extracted thoroughly with ether. The ethereal solution was dried over anhydrous magnesium sulfate, and the ether was removed to leave a residue of 42 g. of strawcolored oil which could not be crystallized. The crude abromo-B-methylb-m-nitrobenzoylaminovalericacid (IX) was not purified but was used directly in the next step (ring-closure)

.

1

51s.

Ethyl 3-Methylpyrrolidine-2-carboxylate (xrv).-A

so-

lution of 27.5 g. of crude l-m-nitrobenzoyl-3-methytpp-

(17) The procedure used here is similar to that followed by Dr. M. Carmack in the addition of ethyl I-pyrrolidine-2-carboxylateto ethyl acrylate.

Feb., 1944

PROOFOF THE STRUCTURE OF RETRONECINE

0 tically Active Ethyl &N-(&Methyl-2-carbethoxypyrroli&l)-propbmte.-A solution of 1.2 g. of the ethyl 3methylpyrrolidine-2-carboxylate prepared from optically active starting material and 4 cc. of ethyl acrylate was refluxed on a steam-bath for twenty hours, protected from moisture. The product boiled at 170-171' (25 mm.); n% 1.4512; yield, 1.5 g. Anal. Calcd. for CIaHaN04: C, 60.69; H, 8.98; N, 5.44. Found: C, 60.88; H,9.16; N, 5.69. Rotation. 0.1038 g. in 2.5 cc. of absolute ethanol at 30' g a v e a ~-1.45'; I , 1; [ala! -34.9' (t0.5'). l-Methyl-7-ketopyrrolizi&e (xvr).-Metallic potassium (2.5 9.) was finely powdered under 20 cc. of dry xylene. The mixture was cooled and 10 cc. of dry benzene was added. With Vigorous stirring, a solution of 6 g. (23 millimoles) of optically inactive ethyl O-N-(3-methyl-2carbethoxypyrrolidy1)-propionateand 10 cc. of dry benzene was added dropwise over a period of thirty-five minutes. The residual diester solution was washed in with an additional 10 cc. of dry benzene and the reaction mixture was heated under reflux on the steam-bath for four and one-half hours, protected from moisture. The mixture was cooled and the excess potassium was decomposed by the addition of 8 cc. of ethanol, followed by 20 cc. of distilled water. The mixture was separated and the benzene-xylene solution was extracted with four 5-cc. portions of water. To the combined aqueous extracts was added 30 cc. of concentrated hydrochloric acid, and the solution was heated under reflux for two and one-half hours on the steam-bath. After the hydrochloric acid had been removed by distillation in vacuo, the dark reddish residue was taken up in 20 cc. of water, basified with saturated sodium hydroxide solution, and extracted five times with ether. The ethereal solution was dried, the ether removed, and the residue distilled under reduced pressure. The optically inactive 1-methyl-7-ketopyrrolizidinewas obtained as a colorless mobile liquid which rapidly darkened; b. p. 96.598' (18 mm.), 79-80' (8 mm.); n% 1.4770; yield, 1.9 g. (59%). \ - - , V I

Anal. Calcd. for C8HlsNO: C, 69.03; H, 9.41. Found: C, 68.91; H, 9.33. Picrate.-Prepared in anhydrous ether and recrystallized from absolute ethanol, the picrate formed yellow, feathery elongated prisms which melted, with decomposition, at 189-190° (cor.). Anal. Calcd. for ClrHlsNdOa: C, 45.65; H, 4.37. Found: C, 45.73; H, 4.49. Metitiodide.-Prepared in and recrystallized from dry benzene the methiodide formed colorless prisms, m. p. 149.5-150.5' (cor.). Anal. Calcd. for G H d N O : C, 38.46; H, 5.73. Found: C, 38.33; H, 5.94. I-Menthydrazone.-A solution of 1.9 g. of optically inactive 1-methyl-7-ketopyrrolizidineand 2.25 g. of I-menthydrazide" in the minimum amount of absolute ethanol (containing 2 g. of fused sodium acetate and 1 g. of glacial acetic acid per 100 cc. of absolute ethanol) was heated under reflux on the steam-bath for twenty-five hours, protected from moisture. After standing a t room temperature, crystals separated. These were collected on a filter, washed with a little absolute ethanol, and recrystallized from ligroin (b. p. 90-110'). After the firSt recrystallization, 160 mg. of colorless elongated prisms was obtained; these melted at 169.5-171' (cor.) and the melting point was unchanged after three more recrystallizations from ligroin. Anal. Calcd. for ClpHsaNaOe: C, 68.02; H, 9.92; N, 12.53. Found: C,68.09; H,9.79; X, 12.64. Rotation. 0.0658 g. made up to 5 cc. with absolute ethanol a t 27' gave QD -0.52'; I , 1; [cu]*'D -39.5' (*0.5'). After the above crystals had been removed by filtration, the ethanolic filtrate was heated and diluted with water. (18) Woodwud, KohmPn and Hprrir, THUr J O - N ~ , (1941).

6% 120

261

When the solution was allowed to cool slowly, 2.7 g. of crystals was obtained and submitted t o a systematic sixstage fractionation from ligroin. The melting points of all of the fractions remained nearly identical and the specific rotations of widely separated fractions also were identical within the precision of measurement. Chromatographic adsorption of a dilute ligroin solution of the l-menthydrazone on activated alumina brought about no recognizable separation of the diastereoisomers. In order to check the assumption that no resolution was being accomplished through this derivative, the most insoluble fraction of Imenthydrazone obtained at the sixth fractional recrystallization stage was converted t o the oxime. Oxime.-Approximately 0.6 g. of the most insoluble portion of I-menthydrazone was retluxed for one hour in 20 cc. of 2 N sulfuric acid. The solution was cooled, extracted with ether and the ether extract discarded. The aqueous solution was made strongly basic with sodium hydroxide and extracted again with ether. The combined ether extracts of the alkaline solution were dried over anhydrous magnesium sulfate and evaporated t o small volume. Addition of excess picric acid in dry ether resulted in the precipitation of yellow crystals. This picrate, after one recrystallization from absolute ethanol, melted at 189-190' (cor.) and gave no depression of melting point when mixed with the picrate of optically inactive 1-methyl-7-ketopyrrolizidine. The picrate was decomposed in hydrochloric acid solution, and the picric acid was removed by extraction with ether. The aqueous s6lution was evaporated to dryness i n vacuo. The residue was redissolved in 3 cc. of water and treated with 3 cc. of 5 N sodium hydroxide and 0.25 g. of hydroxylamine hydrochloride. The oxime was formed and isolated in the same manner as described for the oxime of retr~necanone.~Recrystallization from chloroform-petroleum ether (b. p. 40-60') gave colorless prisms, m. p. 128-130' (cor.). The compound showed no optical activity. -Anal. calcd. for GHIIN~O:C, 62.31; H, 9.15; N, 18.17. Found: C62.54; H,9.16; N, 18.26. Optically Active l-Methyl-7-ketopyrrolizidine.-Cyclization of levorotatory ethyl fi-N-(3-methyl-2-carbethoxypyrrolidy1)-propionate was carried out under the same conditions employed in the formation of the optically inactive 1-methyl-7-ketopyrrolizidine(XVr). A suspension of 0.6 g. of finely powdered potassium in 6 cc. of d r y xylene and 3 cc. of dry benzene was treated with a solution of 1.3 g. of the levorotatory diester and 6 cc. of benzene added dropwise during one-half hour. The aminoketone was obtained as 0.4 g. of light amber-colored oil which was not distilled but was converted directly t o the oxime in the usual manner, by treating with 0.6 g. of hydroxylamine hydrochloride in 6 cc. of 2.5 il.' sodium hydroxide. The oxime was purified by recrystallization from chloroformpetroleum ether (b. p. 40-60'), from which it separated as colorless elongated prisms, m. p. 166-167' (cor.). Anal. Calcd. for C8Hl&O: C, 62.31; H, 9.15; N, 18.17. Found: C, 62.44; H, 9.06; N, 18.08. Rotation. 0.0659 g. made up t o 5 cc. with absolute ethanol at 32' gave UD -1.02'; 1, 1; [ Q ] ~ ~ D-77.3' ( * 1.5"). The melting point was not depressed when the oxime was mixed with varying proportions of authentic retronecanone oxime (reported m. p. 167-168' (cor.)). The crystalline form of the synthetic oxime was identical with that of retronecanone oxime as observed under the microscope. Furthermore, the specific rotation of this synthetic oxime was identical, within the precision of measurement, with that recorded for retronecanone oxime ( [ a ] * ' ~ -76.0' t2.5'). Picrate of Optically Active 1-Methyl-7-ketopyrolizidhe Oxime.-Prepared in anhydrous ether, from the oxime of optically active 1-methyl-7-ketopyrrolizidine and dry picric acid, and recrystallized from absolute ethanol, the picrate formed yellow elongated prisms which melted, with decomposition. at 188-190' (cor.). A m i . ~ c h for . c ~ , H ~ , N ~ oC, ~ :43.86; H, 4.47; N, 18.27. Found: C, 43.97; H, 4.66; N, 18.23.

ROGBRADAMSAND N. J, LEONARD

262

The melting-decomposition point was s o t depressed when the synthetic picrate was mixed with the picrate of retronecanone oxime. Picrolonate of Optically Active l-Methyl-7-ketopyrrolizidine Oxime.-Prepared in absolute ethanol and rccrystallized from ethanol-benzene solution, the picrolonate formed clusters of yellow elongated prisms which melted, with decomposition, at 209-211' (cor.). Anal. Calcd. for ClsHrzNeOs: C, 51.67; H, 5.30; N. 20.09. Found: C, 51.83; H, 5.48; N, 20.02. The melting-decomposition point was not depressed when the picrolonate was mixed with the picrolonate of retronecanone oxime. Retronecanone Picro1onate.-Prepared in and recrystallized from absolute ethanol, the picrolonate formed yellow elongated prisms which melted, with decomposition and foaming, at 195-196' (cor.). Retronecanone Methiodide.-Formed by warming retronecanone with excess methyl iodide in benzene solution and recrystallized from absolute ethanol, the methiodide formed colorless elongated prisms, m. p. NO-161' (cor.). Retronecanone I-Menthydrazone.-An absolute ethanol solution of 0.8 g. of retronecanone and 0.95 g. of l-menthydrazide was heated on the steam-bath for twenty-four hours under reflux. When the reaction mixture was cooled and diluted with water, 0.8 g. of fine elongated prisms separated in a closely adhering mass. After seven recrystallizations from ligroin (b. p. 90-110') the product melted at 175.5-176.5' (cor.). AnaZ. Calcd. for C19H83N1101:C, 68.02; H, 9.92; N, 12.53. Found: C, 68.20; H, 9.95; N, 12.65. Rotation. 0.0655 g. made u p o t o 5 cc. with absolutz ethanol at 28' gave :a -1.09 ; I, 1; [ ( Y ] ~ * D-83.2 (*0.5'); [ M I P 8 D-279 Hydrolysis of Retroneceone I-Menthydrazone.-A representative sample was boded with 20 cc. of 1 N sulfuric acid for one hour. The solution was cooled and extracted with ether, which removed most of the coloration. After the aqueous solution had been made strongly basic, it was thoroughly extracted with ether, the combined ether extracts dried, and the ether solution concentrated to 10 cc. Picric acid in anhydrous ether was added until there was no further yellow precipitation. The solid picrate was collected on a filter and recrystallized from ethanol, from which it separated as fine elongated prisms, m. p. 195' (cor.), with decomposition. There was no depression of melting point when the sample was mixed with retronecanone picrate. Picrate of Retronecanone Oxime.-Formed in anhydrous ether from picric acid and retronecanone oxime3 and recrvstallized from absolute ethanol, the picrate formed yellow elongated prisms which melted, with decomposition, a t 188-190' (cor.). Picrolonate of Retronecanone Oxime.-Prepared in absolute ethanol and recrystallized from ethanol-benzene, the picrolonate formed clusters of yellow elongated prisms which melted, with decomposition, a t 209-2211' (cor.). Acetylretronecanol Methiodide.-This derivative of retronecanol was prepared according to the method of Barger? Seshadri, Watt and Yabuta,l* but the stout color; less prisms from acetone were found to melt a t 215-216 (cor.) instead of 207-208' as reported. Rotation. 0.0709 g. made up to 5 cc. with absolute methanol a t 28" gave ~ Y D-1.24'; 1, 1 ; [ c Y ] * ~ D-87.4" ( * 1.0"). Reduction of Retronecanone (a) Neutral Solution.-A solution of 3GG mg of retronecanone in 20 cc. of absolute ethanol was hydrogenated at normal pressure over 50 mg. of platinum oxide catalyst. The theoretical volume of hydrogen wa9 absorbed in one hour. The reduction product was not obtained crystalline but the impure oil obtained exhibited a specific rotation in D (c = 1.266). This differs absolute ethanol of [ c Y ] ~ ~-9.5" widely from that reported for retronecanol (-91.1').13 The methiodide, obtained crystalline a t first from acetone solution, was extremely hygroscopic (unlike retronecanol rnethiodide) and could not be recrystallized.

.

Vol. 66

The picrate, formed in ether solution and recrystallized from absolute ethanol, separated as yellow elongated prisms which melted, with decomposition, at 218-219' (cor.). Retronecanol picrate melts, with decomposition, at 210' (cor.).la Anal. Calcd. for CtrHlsN4Os: C, 45.41; H, 4.89 Found: C, 45.51; H, 4.96 The reduction product of retronecanone was acetylated by heating under reflux on the steam-bath for four hours with excess acetyl chloride. The acetyl chloride was removed in uucuo, leaving a brown semicrystalline residue which was dissolved in 5 cc. of water. The aqueous solution was made alkaline with potassium carbonate and the acetylated base was extracted with ether. The ethereal solution was dried, decolorized and the ether removed. The residue did not crystallize, but the methiodide was obtained crystalline from benzene. After repeated fractional crystallizntion from acetone, the least-soluble fraction, obtained as colorless elongated prisms, possessed a maximum constant melting point of 210-212' (cor.). No other methiodide was obtained pure in the fractional crystallization process. Anal. Calcd. for CI~HJNOI: C, 40.62; H, 6.20 Found: C, 40.79; H, 6.37. Rotation. 0.0234 g. made up t o 2.5 cc. with absolute methanol a t 28' gave r r ~$0.07'; 1, 1; [ ~ ] * S D +7.5' (*1.5'). The rotation differs widely from that of acetylretronecanol methiodide (-87.4'), and a melting point of a mixture of the two methiodides is depressed to 135-190'. (b) Acid Solution.-Reduction of retronecanone over platinum oxide catalyst was carried out in 20 cc. of absolute ethanol containing 1 cc. of concentrated hydrochloric acid. The reduction was complete after nine hours. The reduced free base was liberated but was not obtained crystalline. The methiodide, obtained crystalline at first from acetone solution, was hygroscopic and could not be recrystallized. The picrate, formed in and recrystallized from absolute ethanol, separated as yellow platelets which melted, with decomposition, a t 230-232' (cor.). The picrate differed from the picrate of the reduction product of retronecanone in neutral solution, and also from retronecanol picrate and oxyheliotridane picrate." Anal. Calcd. for C I ~ H I ~ N ~N, O 15.13. ~: Found: N, 15.23. Carbon and hydrogen values are not reported since the analytical sample exploded. Acetylation of the reduced product was carried out as before and the picrate of the acetylated reduction product was formed in anhydrous ether and recrystallized from absolute ethanol: yellow elongated prisms, m. p. 178.5179.5' (cor.). Anal. Calcd. for ClsHt&aO~: C, 46.60; H, 4.89. Found: C, 46.64; H , 4 91. The methiodide of the acetylated reduction product was formed in benzene and subjected to repeated fractional crystallization from acetone. The least-soluble rnethiodide fraction was finally obtained as colorless elongated prisms, m. p. 210.5-211.5' (cor.). This was identical with the methiodide of the acetylated reduction product of retronecanone in neutral solution and gave no melting point depression when mixed with it. Mixing with acetylretronecanol methiodide again produced a sharp depression of melting point. N o other methiodide could be obtained pure in the fractional crystallization process. Reduction of I-Methyl-7-ketopyrrolizidhe: 1-Methyl7-hydroxypyrrolizidme.-Hydrogenation of 520 mg. of optically inactive 1-methyl-7-ketopyrrolizidinein 20 cc. of absolute ethanol was carried out a t atmospheric pressure in the presence of 50 mg. of platinum oxide catalyst. The theoretical volume of hydrogen was taken up in one-half hour, after which the ethanol solution was filtered free of catalyst and used directly for the formation of derivatives. Neither tbe hydrochloride nor the methiodide was ob-

THEPREPARATION OF SOME SUBSTITUTED PIPERAZINES

Feb., 1944

263

tained pure because both were very hygroscopic. The picrate and picrolonate were obtained as well-characterized crystalline salts. Picrate.-Prepared in and recrystallized from absolute ethanol, the picrate formed yellow elongated prisms which softened at 210' and melted, with decomposition, at 218.5219.6' (cor.). Anal. Calcd. for C14HIIN40~:C, 45.41; H, 4.89. Found: C, 45.57; H, 5.06. Picro1onate.-Prepared in and recrystallized from absolute ethanol, the picrolonate formed yellow elongated prisms, m. p. 182.5-183.5' (cor.). It is interesting that the melting point of the racemic 1-methyl-7-hydroxypyrrolizidine picrolonate is almost the same as that of retronecanol picrolonate, m. p. 184-185O.10 These picrolonates,of course, are not identical. Anal. Calcd. for ClgHlllN~O~: C, 53.33; H, 5.72; N, 17.28. Found: C, 53.26; H, 5.71; N, 17.32.

1-menthydrazone of retronecanone in melting point and rotation. 2. Optically active ethyl 3-methylpyrrolidine2-carboxylate was converted to the corresponding optically active 1-methyl-7-ketopyrrolizidine by a similar series of reactions. By treatment of the reaction mixture, with hydroxylamine, a pure optically-active 1 - methyl - 7 - ketopyrrolizidine oxime resulted, which was identical with retronecanone oxime in melting point and rotation. The identity was verified further by the preparation of identical picrates and picrolonates. 3. It has been proved that the keto group in retronecanone is in the 7-position. From this it may be deduced that the secondary hydroxyl group in retronecanol, retronecine, platynecine, Summary desoxyretronecine, heliotridine, and oxyheliotri1. 1-Methyl-7-ketopyrrolizidine has been syn- dane is in the 7-position. 4. It has been shown that in the reduction of thesized through ethyl 3-methylpyrrolidine-2desoxyretronecine, retronecine, and esters of recarboxylate (obtained from ?-picoline) by addition to ethyl acrylate followed by a Dieckmann tronecine to give retronecanol, asymmetric reduccyclization and hydrolysis. The final product tion has taken plaec with the exclusive formation probably was a mixture of two racemates. By of one stereochemical configuration of the CI-CHs. 5. Reduction of retronecanone in neutral and condensation with E-menthydrazide a single pure 1 - methyl - 7 ketopyrrolizidine - E - menthy- acid solutions gave a mixture of corresponding drazone was isolated. This was different from the hydroxy compounds, from which retronecanol could not be isolated. (19) Konovdova and Orekhov, ,Bull. SOC. chim., [5] 4, 1286

-

URBANA, ILLINOIS

(1937).

RECEIVED NOVEMBER 8, 1943

[CONTRIBUTION FROM THE BURROUGHS WELLCOME & CO. u. s. A. EXPERIMENTAL RESEARCH LABORATORIES]

The Preparation of N-Mono-substituted and Unsymmetrically Disubstituted Piperazines BY RICHARD BALTZLY, JOHANNES

s. BUCK,EMILLORZAND WILLIAMSCHON

The direct alkylation of piperazine produces a reaction, or if the acylating or alkylating agent mixture of substitution products from which the were to form a second liquid phase, thus tending to isolation of mono-substituted compounds is diffi- extract preferentially mono-substituted piperazine cult or impossible when the substituting group is already present. relatively small. Because of this the belief has The yields here recorded compare favcrably arisen that mono-substitution products are not with those claimed for more involved procedures. formed, although examination of reported experi- The argument by which Moorela was led to operments shows only that they were not isolated. A ate a t low PH is probably valid. The second similar situation has prevailed in regard to acyla- benzylation recorded in Table I, run at pH about tion, although even mono-acetylpiperazine can be 3.5 gave a proportion of mono-benzyl to dibenzyl obtained easily by logical manipulation. piperazine of about 5-1. Similarly a proportion of In Table I are shown the yields of mono- and di- 2.7-1 was observed in the carbethoxylation, where substituted piperazines from a number of reac- a low pH must have existed toward the end of the tions, together with the reaction conditions. It reaction. As a preparative method with rnoderseems indicated that in the ideal case where the ately priced chemicals, however, the buffering entire reaction takes place in one homogeneous procedures seem unprofitable. phase, equimolecular proportions of reactants reFrom the mono-substituted piperazines that sult in the 1:2: 1 proportions of piperazine, mono- were thus rendered available a number of unsubstituted piperazine and di-substitution product symmetricallydisubstituted derivatives have been predicted on general considerations. prepared. These are characterized in Table 11. Major deviations from these proportions would (1) (a) Moore, Boyle and Thorn, J . Chcm. SOC.,39 (1929); (b) be expected if a salt of piperazine itself (such as the Jacobi, Be?., 66, 113 (1933). The methods descrihed in these papers are two-phase acylations at low pH. The conditions are difficult to dihydrochloride which is sparingly soluble in reproduce exactly and it is doubtful if yields significantly above 60% absolute ethanol) were to precipitate during the should be expected from them.